Review Sentiment Analysis and Verification Status Classification

The below paper is the final term project for the fall 2024 CIS 9665 - Applied Natural Language Processing course at Baruch College.

Using a dataset of categorized Amazon reviews hosted by the University of California - San Diego Rady School of Management, our team created a binary classification model to predict customer review verification status.
Given the increasing importance of online marketplaces and the value of customer trust, we sought to create a solution with a wide range of applications.
Implementing several natural language processing techniques such as sentiment analysis as well as classifying review phrasing and -contents, we built a Random Forest classification achieving an accuracy of 86%. Since the model was created on a subset of the data, the preprocessing steps and model creation can be easily reproduced and applied in academic and commercial settings.

Team member names have been redacted for privacy considerations.